I disagree a bit here, and not because of the price estimation, but because I think there is some space for higher transaction fees than today.

In 2017 the average transaction fee miner income was well above 1 BTC/block and Bitcoin continued to work without flaws for most of the year. In December it reached 10 BTC/block at one point.

I think 10 BTC/block (that was about 30 USD for a standard transaction of ~200 bytes back then) wasn't sustainable before Segwit. 30 USD for a standard transaction is clearly too high, it would allow Bitcoin to work only as a settlement layer, and thus I don't believe this level can be held because of altcoin competition. But after Segwit - where capacity has approximately doubled - I consider the 10 BTC/block value a realistic upper boundary. At today's prices, 10 BTC/block means an average transaction price (for the 180 byte standard transaction) of 6 USD. This is high but for Bitcoin's killer feature - international transactions - it's still low. It would also allow the creation of LN channels for less than 20 USD and possibly less with channel factories.

For the long term I think, if we stay with approximately the current price (let's say between $3000 and $10000) a miners' fee income of 1 BTC/block (like in mid-2019) is very realistic and 3 to 5 BTC/block (which already would be more than the reward after 2025) not unlikely. That means that we basically don't need to care about the reward as long as blocks are well filled (>80-90% of capacity). If price increases, then the numbers will obviously even be better for the miners.